Geometry of enolization using a bifunctional cyclodextrin-based catalyst

make it possible to explore the geometric preferences of other ... compounds. (4) Since the catalyst concentration is well in excess of that needed fo...
0 downloads 0 Views 156KB Size
J. Am. Chem. SOC.1993, 115, 10988-10989

10988

Communications to the Editor Geometry of Enolization Using a Bifunctional Cyclodextrin-Based Catalyst 3.0

Ronald Breslow' and Alan Graff Department of Chemistry, Columbia University New York, New York 10027 Received August 27, 1993

e

f 2.5

We have described1 a set of 8-cyclodextrin (cycloheptaamylose) bisimidazoles in which the catalytic groups are located on the C-6 carbons of neighboring glucose units (AB), or separated by one unit (AC) or two units (AD). Thus the attachment points are respectively 51°, 103O, and 154O apart around the seven-unit cyclodextrin circle into which substrates can bind. We saw that the AB compound was the best catalyst for hydrolyzing a bound cyclic phosphate ester,' consistent with a mechanism we proposed2 for that process. The availability of this set of catalysts should make it possible to explore the geometric preferences of other bifunctionally catalyzed processes and to mimic some enzymatic catalyses. We now wish to describe such a study with respect to the enolization of a bound ketone substrate. The substrate wasp-tert-butylacetophenone(1). We monitored the exchange of deuterium into the methyl group of 1 by GC/MS using CI (CH4), examining3 the peaks at [M + 291, [M + 301, [ M + 311, and [ M 321. Deuterium exchange measures the rate of enolization assuming that BH+ of the catalyst exchanges rapidly with D2O relative to the reketonization rate; if not, the enolization rate is even faster than deuteration. The catalyst was at 6 mM and the substrate at 2 mM. and the medium was phosphate buffer at 170 mM in D2O with 14% CD30DU4The pH6 did not change within 0.05 units during the experiments, which were performed at 35 OC. The results of experiments at pH 6.2 (where the imidazole groups should be ca. 50% protonated) are shown in Figure 1. Under our conditions, there is no detectable deuterium exchange catalyzed by buffer alone (or by a solution of 8-cyclodextrin and imidazole), but there is significant catalysis by the three cyclodextrin bisimidazoles and also by cyclodextrin monoimidazole. As Figure 1 shows, the AB and AC bisimidazoles are not significantly better than the monoimidazole catalyst, but the rate with the AD isomer is faster. This indicates that the AD isomer is a true bifunctional catalyst, a conclusion supported by the pH vs rate constant profile in Figure 2. By molecular modeling, all three of the bifunctional isomers can reach the carbonyl oxygen and the C-H bond being broken in a complex with substrate 1. However, as shown in Figure 3, the stereoelectronics are different. With AB, the imidazole can reach the H approaching from the side toward the carbonyl group

0*

5 r

i z

2.0

1.5

+

(1) Anslyn, E.; Breslow, R. J. Am. Chem. SOC.1989, 111, 5912-5973. (2) Breslow, R. Acc. Chem. Res. 1991, 24, 317-324. ] were more reliable than were the [M + 11 (3) These [M C ~ H Jpeaks peaks. [M + 291 results from I-Ha, while [M + 301 results from I-HzD and also 1-H3.I3C. To follow the exchange of proton, [M + 301 was corrected for the I3C abundance seen in the starting 1, and similar corrections were applied to [M + 311 and [M + 321 to obtain the relative amount of the D2 and D3 compounds. (4) Since the catalyst concentration is well in excess of that needed for saturation binding of a rerf-butylphenyl group into j3-cyclodextrin in water containing this concentration of ethanol5 and ethanol is more hydrophobic than methanol, these must correspond to saturation binding conditions. ( 5 ) Breslow, R.; Halfon, S. Proc. Narl. Acad. Sci. U.S.A.1992,89,69166918. (6) Measured with a normal pH electrode uncalibrated for our medium. Thus the relative values have meaning, not the absolute values.

+

0002-786319311515-10988$04.00/0

22

= E

20

-

-

181s:

I

19

-E.

1t

.

14-

12

'-

l:;

5: 4 -

2 -

or

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 . . . . , . . . . , . . . . ~

Communications to the Editor

J . Am. Chem. SOC.,Vol. 115, No. 23, 1993

(AC)

(AB)

Y

Figure 3. Catalyzed exchange reaction.

if the ImH* proton is placed directly on an oxygen unsharedelectron pair with 0,H, and N in a line. With AC, the imidazole can reach the H in line with the C-H bond, while with AD it reaches the H from a direction somewhat behind the C-H bond (7) For an example of bifunctional catalysis of enolization in wet CHCl,, cf.: Wolfe, J.; Muehldorf, A.; Rebek, J., Jr. J . Am. Chem. SOC.1991, 113, 1453-1454.

10989

(away from the carbonyl group). Since the electrons of the C-H bond move toward the carbonyl group on conversion to the en012 as the CH3 carbon rehybridizes, a preference for such a nonlinear backside approach seems reasonable. This is probably the geometry for base abstraction of the H in all the cases, with the difference being whether the acid catalyst group can get in the right position to assist. It will be interesting to see whether our catalysts-the first to combine a hydrophobic binding site with two acid-base catalytic groups in a well-defined position and to operate in aqueous solution7-can selectively promote other reactions involving enolization. The set of isomeric catalysts will also let us explore the Occurrences and geometric preferences of other acid-base bifunctional catalyses. Meanwhile, the clear contrast with the geometric preference for bifunctional catalysis of phosphate ester hydrolysis, in which the AB isomer was the best, supports our interpretation of these preferences in terms of the mechanism and optimal geometry of the reactions involved. Acknowledgment. This work was supported by the National Institutes of Health and the Office of Naval Research. A.G. is the holder of a Feodor Lynen Fellowship from the Humboldt Foundation.